Mechano-pharmacology of T cell receptor specificity

As with many receptors, closely related T cell receptor (TCR) ligands can be characterized functionally as agonists, non-agonists and antagonists; changes to a single methylene group in the peptide bound to major histocompatibility complex (pMHC) on antigen-presenting cells determine these outcomes in combination with physical forces that help separate agonists from non-agonists. Qin et al. in Cell Research show that the co-receptor CD8 also contributes to this mechano-pharmacology, explaining the large dynamic range of the TCR‒pMHC‒CD8 system and dangers of self-peptide cross-reactivity inherent to engineered high-affinity TCR.

Over the last three decades, high-affinity T cell receptor (TCR) engineering has rapidly advanced utilizing phage, yeast, and T cell display systems^1,2,3 to promote both diagnostic and therapeutic effects. These high-affinity TCRs have up to 1000-fold stronger binding in solution (3 dimension, 3D) for peptide-major histocompatibility complex (pMHC) than natural TCR, which can lead to better anti-tumor efficacy.⁴ However, early efforts with high-affinity TCRs were limited by cross-reactivity with tissue-specific self-antigens, which are difficult to fully exclude. A greater understanding of both the on-target reactivity and off-target cross-reactivity could improve the design of therapeutic high-affinity TCRs.

TCR naturally interacts with pMHC at membrane interfaces, such that two-dimensional (2D) affinities of the TCR‒pMHC interaction in situ correlate better than 3D affinities with TCRs’ potency.^5,6 The evolution of such 2D measurements later incorporated mechanical force as a key factor to regulate the strength of TCR‒pMHC bimolecular^7,8 and TCR‒pMHC‒CD8 trimolecular⁹ interactions at membrane interfaces. Evidence suggests that forming catch bonds that are transiently strengthened under 10 pN forces is a characteristic of agonistic pMHC. However, how mechanical force regulates the engineered high-affinity TCR‒pMHC‒CD8 trimolecular interactions and their self-peptide cross-reactivity had been unknown.

Qin et al.¹⁰ employ a broad range of powerful tools, including mutagenesis, structural and functional analysis, single-molecule biomechanical manipulation, MD simulation, and mathematical modeling, to demonstrate that the TCR specificity is collectively regulated by the TCR‒pMHC‒CD8 trimolecular catch bonds under mechanical force, and provide an explanation for how the engineered high-affinity TCRs acquire CD8-dependent cross-reactivity with self-peptides complexed with MHC even as there direct recognition of the agonist can become CD8 independent (Fig. 1).

A natural TCR forms catch bonds with R4-H2-K^b under an optimal force at 10 pN, which is further enhanced by the right participation of CD8 with molecule conformational changes and rotations. However, it forms slip bonds with L4-H2-K^b, and CD8 can no longer enhance this interaction. In contrast, an engineered high-affinity TCR forms catch bonds with both R4-H2-K^b and L4-H2-K^b with false contribution from CD8. The mechanical force-regulated TCR‒pMHC‒CD8 trimolecular interactions explain why engineered high-affinity TCRs have cross-reactivity with non-agonist peptides and induce false T cell responses.

Full size image

Qin et al. found that CD8 was required for mouse 2C TCR to activate T cells, specifically, the mouse T-cell hybridoma lines, in response to the agonistic (arginine at position 4 = R4) peptide bound to MHC class I protein H2-K^b. In contrast, the non-agonistic (leucine at position 4 = L4) mutant of the same peptide bound to H2-K^b with or without CD8 did not lead to T cell activation. Two high-affinity versions of the 2C TCR, designated as m33 and m67 TCRs, were expressed in T cells, which were activated by R4-H2-K^b without or with CD8, and L4-H2-K^b with CD8, but not without. Paradoxically, R4-H2-K^b was more potent at activating T cells with the 2C TCR than with either of the affinity-enhanced variants. These findings raised a number of questions — why don’t the affinity-enhanced TCRs use CD8 properly, and how does the same TCR then use CD8 to recognize the non-agonist ligand?

Qin et al. first showed that 2C TCR forms catch bonds with R4-H2-K^b over a wide force range with an optimal force at 10 pN and bond lifetime of 0.25 s. What’s more, CD8 further prolongs this bond lifetime by 2-fold. However, under the same experimental conditions, 2C TCR forms slip bonds with L4-H2-K^b, and CD8 could not enhance this interaction (Fig. 1 left and yellow lines). In contrast, m33 and m67 TCRs form catch bonds with both R4-H2-K^b and L4-H2-K^b with much longer bond lifetimes than parental 2C TCR (Fig. 1 right and red and blue lines). These data clearly demonstrate that m33 and m67 TCRs interact more tightly with agonist peptide-MHC, but meanwhile show mechano-pharmacological enhancement with the non-agonist L4-H2-K^b, which could explain how T cells with m33 and m67 TCRs respond to L4-H2-K^b stimulation, but could not explain why the affinity-enhanced TCRs do not benefit from CD8 when recognizing R4-H2-K^b.

Qin et al. then used constant-velocity steered molecular dynamics (cv-SMD) at atomic resolution to demonstrate that the CD8’s participation in 2C TCR‒R4-H2-K^b interaction is regulated by mechanical force with contributions from the conformational changes and rotations of MHC and CD8 molecules. Under the regulation of mechanical force, TCR‒pMHC interactions that are either too weak (2C TCR‒L4-H2-K^b) or too tight (m33 TCR‒R4-H2-K^b, m67 TCR‒R4-H2-K^b) attenuate the contribution of CD8 to this system. To broaden the scope of the research conclusions, Qin et al. also validate the force-regulated catch bonds of engineered high-affinity human TCR (MAG-IC3-TCR) with agonist peptide MAG-A3 bound to human MHC class I HLA-A*01:01 both in the presence and absence of CD8 participation. However, they find that MAG-IC3-TCR forms a catch bond with the cardiac muscle-associated Titin-HLA-A*01:01, providing a mechano-pharmacological mechanism for off-target cross-reactivity of an affinity-enhanced TCR. Finally, based on their characterized parameters, the authors provide TCR‒pMHC kinetics-function maps with mathematical modeling, to guide engineering of TCR with high efficacy and less off-target risks, which is of both scientific and clinical significance.

Overall, the findings in this study demonstrate that the antigen-driven function and specificity of T cells are not determined by the 3D affinity of TCR to pMHC, but by the mechanical force-regulated TCR‒pMHC‒CD8 trimolecular interactions at 2D interfaces. This study contributes not only to our fundamental understanding of basic biology of how T cells achieve sensitivity and specificity in antigen recognition but also to the translational research aiming at engineering T cells for adoptive cell therapy for cancer and other diseases.